The Autonomy Paradox: Trust and Control in the Agentic Age

Moderator: Jules Polotnetsky, CIPP/US, CEO, Future of Privacy Forum

Christoph Bausewein, CIPP/E, CIPT, Assistant General Counsel, Data Protection and Policy, CrowdStrike 

Michel Combot, Chief Technology, Innovation and AI Officer, CNIL

Lanah Kammouried Donnelly, Global Head of Privacy Policy, Google

As artificial intelligence products evolve from generative tools into autonomous agents capable of planning and executing complex tasks, their core value proposition — autonomy — presents a fundamental paradox. This paradox exists because an agent's value is directly proportional to its autonomous access to data, the very capability that creates new vectors for privacy harms and security failures. Navigating this trade-off is the central challenge for developers, regulators and practitioners. How do we strike the right balance? How do we increase the immense utility of these systems while ensuring robust privacy and security safeguards?

This panel will explore the premise that for AI agents, data protection and cybersecurity are two sides of the same coin and identify the core principles essential for responsible innovation. The session will then bring next-generation research concepts — such as "contextual integrity" — to the forefront as potential frameworks to advance the privacy frontier.

What you will learn:

• How data protection principles and cybersecurity architectures — such as "human in the loop" and the "zero trust framework" — converge for AI agents.   
• Responsible practices in the development of agentic AI systems (e.g., gradual rollouts, safety testing).  
• Insights into next-generation concepts like 'contextual integrity' as potential frameworks to advance the privacy frontier.

Sponsored by: Google